U.S. President Donald Trump said he viewed AI company Anthropic as a possible national security threat just a week ago, but said the situation has improved after the company responded to government concerns.
Trump made the remarks during an interview for The Axios Show when he was asked whether he considered Anthropic or its CEO, Dario Amodei, a threat to U.S. national security. Trump replied, “Well, not now, but a week ago, maybe,” suggesting that his concerns were much higher only days earlier, as per the interview with Axios.
The issue is important because Anthropic found itself at the center of a major government crackdown related to AI security. Trump said his view changed after meeting Amodei during the recent G7 Summit.
Trump changes view on Anthropic
After the meeting, Trump described Amodei as “nice” and “smart,” showing a much friendlier attitude toward the AI executive. Trump said Anthropic quickly responded after the administration raised concerns about potential risks linked to its technology. “He responded to us very quickly,” Trump said, referring to Amodei and Anthropic’s leadership, as per interview with Axios.
Trump warned that the issue was a “tremendous liability” and suggested it was serious enough to bring major legal consequences if not addressed properly. He added that Anthropic handled the situation “very responsibly,” indicating satisfaction with the company’s response. The dispute began after the Trump administration imposed new restrictions on access to Anthropic’s most advanced AI models.
Also read: End of an era: Trump becomes last president to fly on iconic Air Force One as legendary plane retires
Under those restrictions, countries outside the US and foreign nationals inside the US were blocked from accessing some of Anthropic’s most advanced systems. The administration’s concerns grew after receiving a report from Amazon that identified a vulnerability involving Anthropic’s technology.
Anthropic and government talks
The disagreement led to technical discussions between government officials and Anthropic representatives in Washington. Since then, both sides have reportedly started working together on standards to test and evaluate AI “jailbreaks”, as per the report by Axios. AI jailbreaks are methods used to bypass safety rules and restrictions built into AI systems.
Despite the improved relationship, Trump did not completely rule out stronger government action if needed. He said he still has the option of using powers under the Defense Production Act, a law that gives the president broad authority during national security situations. Trump said, “I have the power to use a lot of things,” when asked about the possibility of invoking the Defense Production Act, as noted in the interview with Axios. However, he also suggested that such action may no longer be necessary because the situation appears to be improving.
Trump keeps pressure on AI firm
Trump claimed that concerns about Anthropic were first brought to the government by a competitor that also owned part of the company. He said the group that raised the alarm was worried about Anthropic’s actions and potential risks.
Trump specifically pointed to Amazon’s concerns and said those warnings helped trigger government scrutiny of the company. Even so, Trump repeated that Anthropic’s behavior has been “very responsible” since the issue was raised.
In response, Anthropic released a statement thanking the administration for working with the company to resolve the matter. The company said it appreciated the ongoing partnership with the government and wanted a quick solution to the dispute.
